This proposed project will combine U.S. Army Health Risk Appraisal (HRA) data, including self-reports of drinking and other risk taking behaviors, with an existing computerized health outcomes database. HRA data will be linked by social security numbers to hospitalization, death, accident and disability data. The purpose is to quantify the relationship between alcohol use and actual injuries and to identify sub-groups of men and women particularly at risk because of unhealthy drinking behaviors. 8 HRA items, including the 4 which comprise the CAGE, as well as items measuring typical drinks consumed in a week, the use of alcohol while driving or riding with a drinking driver, and peer and personal perceptions about have a drinking problem, will be used to assess drinking behaviors. Approximately 140,000 active duty Army personnel now take the HRA each year. Records for 1989-1994 are available in computerized format providing a total of about 560,000 records available for analyses. Approximately 80 percent of these are first-time takers and the remaining 20 percent of the records represent repeat-takers. Thus, about 500,000 unique responders may be identified and evaluated with respect to their drinking practices and injury experience. The 20 percent repeat responses may be used to evaluate changes in drinking practices and the association between improved (or deteriorating) drinking practices and risk of injury (nested case-control study). The HRA contains items on other risk taking behaviors, such as speeding, use of safety belts, smoking, diet, exercise and weight. Thus, a number of potential confounders of the alcohol injury relationship can be statistically controlled. The potentially synergistic influence of drinking and other factors will also be evaluated through testing of interaction terms. the linked data base contains information on demographic characteristics (gender, age, marital status, number of children, race, occupation, hazardous duty pay). In addition to personnel data, the linked database contains all hospitalization records for active duty army (including ICD-9 CM coded nature of injury, a military code for external cause (e-code), length of stay, case-mix index indicating resources consumed, HCFA and CHAMPUS DRG cost codes, and readmissions); deaths (date, circumstance, location and cause); time-loss injuries reported to the safety Center (activity at time of event, type and cause of injury, severity and body part injured, protective equipment in use, drug and alcohol use, environmental conditions); and disability data (diagnoses and percent disability). This rich source of injury and risk taking behavioral information is unprecedented in the civilian world. SAS will be used to conduct descriptive analyses (chi-square, ANOVA) and multivariate models (logistic and regular regression). Results will be used to identify high risk groups, estimate costs of drinking in the Army, and develop targeted intervention efforts.